LM394 on my Minelab PCB

[ghaj]

Hi Ghaj,

It would be nice to find fets that did have real low C, but they must also have very low on resistance. Any resistance on the input path will create noise(resistor noise) which ends up being the slight or not so slight chatter/warble you can hear on the threshold. Carl mentioned in another thread that 1k input resistance added something like 4 to 5nV noise at the input stage. This difference is very obvious when analysed. So if we have a low R input path, say 5ohms, the noise added is very little, not sure on the exact figure. Then it is the noise floor of the preamp that is dominant in the system, which in this case(ad797) is about .9nV.

Cheers Mick

Thanks Mick. As I have got from your answer, JFETs must have high Ron value. Unfortunately there is no information about Ron in JFETs datasheet.
I have another question. As I have understood by my experiments, the philosophy of using two cascade MOSFETs as input analog switch, and not just one high voltage MOSFET, is to eliminate undesired effect of the intrinsic body diode of the VN2410L, and eventually any other MOSFET used as HV switch. As I have tested, if the P-channel is not placed there, during loop on-state and when the MOSFET is in off state, the body diode of HV MOSFET starts to conduct and lowers output voltage of the switch. Have I got it or there is another reason for this MOSFET arrangement? If I have got it, then I think it is possible to use two cascade N-channel MOSFETs as input FETs, instead of one N-channel followed by one p_Channel. Of course, the second N-Channel must be arranged in reverse direction. This way, we can choose the second MOSFET from a broad range of medium voltage (up to 60V) small signal N-channel MOSFETs with very good Ron and C characteristics. As I have discovered, N-channel MOSFETs are more available and versatile than P-channel ones and they have better switching characteristics in contrast to similar rating P-channel ones.
Thanks for your answer.

[Mechanic]

Hi Ghaj,

It would be nice to find fets that did have real low C, but they must also have very low on resistance. Any resistance on the input path will create noise(resistor noise) which ends up being the slight or not so slight chatter/warble you can hear on the threshold. Carl mentioned in another thread that 1k input resistance added something like 4 to 5nV noise at the input stage. This difference is very obvious when analysed. So if we have a low R input path, say 5ohms, the noise added is very little, not sure on the exact figure. Then it is the noise floor of the preamp that is dominant in the system, which in this case(ad797) is about .9nV.

Cheers Mick

[ghaj]

You are true gentlemen and scholars. With all components now identified, we now have a blue print by which to modify/repair our detectors. Thank you

I made the corresponding updates to the component photo.

Dear Mario,
Have you tried BS107 or BS107A as repalcement for 2SK216? An other question: Have you ever tried to replace P-channel FET with some JFETs? as I have seen in some datasheets JFETs have very very low capacitance (see 2N5460 datasheet for example).

[mario]

I think the CD parts are actually a bit better at handling the faster switching speeds. As good as those specs are for the MAX4066, it can not compete with the faster switching speeds (125ns max vs 40ns max) of the Ti parts.

Those MAX switches (HC), seem to handle larger currents better (30-100mA Vs 10mA) though.

I also installed that BCP53T1G. I can not confirm any changes, but they seem to have over all much lower capacitance than the stock BC369s so I left it installed. They seem to come in a different package, but install very nicely. So far nothing has gotten warm.

[Mechanic]

G'day Mario,

Generally the 74hc or hc labelled parts are for 5v areas, like around the eeprom. The CD prefix denotes that it is for cmos voltage levels 10v, +-5v.

Cheers Mick

[mario]

I had to update the component photo.

The MAX4066 is not a compatible part. Only mixers identified with the prefix "CD" will be direct replacements. The "HC" prefix identifies them as electrical switches. So, if you install a non "HC" part instead of the appropriate replacement, your detector might exhibit ground balancing issues. It wasn't until I started paying closer attention to system timings that I noticed this fault.

A notable replacement for the CD4066 mixer is: CD4066BMTE4, by Ti (get them while they're still available at mouser).



Also, a worth while replacement for BC369 is BCP53T1G.

And last but not least, a replacement for TLE2071, OPA140AID. You can get that at mouser also (while supplies last).

[mario]

You are true gentlemen and scholars. With all components now identified, we now have a blue print by which to modify/repair our detectors. Thank you

I made the corresponding updates to the component photo.

[Woody.au]

Zed is on the money, just scoped it and replaced it with one of the same and the detector still runs.

[ZED]

Looks like chip 9 connects to the 4040 chip (going by the schema) if thats the case it will be a 74hc04 chip.

Pin 10 being a clock input on the 4040 it will probably have a 2 mhz signal going to it from the 74hc04
output pins 2 and 6.
Look for a 2 meg signal using your scope,set your probe to "times 10" otherwise you may load the cct up to much and not see a signal .

Cheers Zed.....

[Woody.au]

I have the same detector but the number is blasted off... I know it is HC74????
The 2200 I have still retains the date code and the only one that fits is below.


PHILIPS 74HC10D A09900Q
HNN9535 C

Scrape the paint off the chip, they may not of obliterated the date code.

[Mechanic]

Hi Woody,

Say you wouldn't be able to put the last piece in the jigsaw puzzle by any chance? The part marked ?9? between the 74hc00 and the 4040?

Cheers Mick

Ps, imagine what could be done with a 4500 with a dedicated high speed micro for timing selection and the existing micro for ground balance and audio filtering. It could be updated with the latest timings out of the 5000

[Woody.au]

YOU GUYS ARE DANGEROUS!!!!!

[mario]

Hi, Mick

Those are nice specs!

Unfortunately, I haven't had time to put everything thing together just yet. I want to take advantage of getting values of all those components hidden behind those caps. I'm going to re-do the portion of both photos without those caps and write component values in a more visible manner. Those caps occupy much room and are hiding a good deal of components . When I get that done, we should get a better idea about their placement.

Another thing that is taking a lot of my time is VMLAB. I finally got it running, emulating my target hardware (Atmega128 with LCD and keys)! -It took me days since the documentation using GCC is some what poor...

Also, I need to still do my small board with the frequency switch...

I'll talk to you later, so much to do...

Mario

[Mechanic]

Hi Mario,

The specs of the opa4140 are fantastic, I can't find anything that is close! My only problem is I cannot find a supplier for them here in Oz I realise the pinout for my detector is different, however there are always ways around that!

I'll see how I go obtaining one or 5

Cheers Mick

[Mechanic]

Hi Mario,

This is an LC oscillator. You see the big resistor just above the little board with the diodes on it? It is actually an inductor. So the combination of the inductor with the capacitors and the smd fet next to the unknown ic form an oscillator cct.

I have found that dedicated vco ic's seem to have too much jitter and cause the detector to be less sensitive.

Cheers Mick